1. Field of the Invention
The present invention relates generally to the field of downhole sampling of hydrocarbons and in particular to downhole and onsite surface high resolution spectroscopy of hydrocarbon samples for measurement and estimation of physical and chemical properties of fluid from a downhole formation before, during or after sample capture in a sample chamber.
2. Background Information
In the oil and gas industry, formation testing tools have been used for monitoring formation pressures along a wellbore, obtaining formation fluid samples from the wellbore and predicting performance of reservoirs around the wellbore. Such formation testing tools typically contain an elongated body having an elastomeric packer that is pressed against the zone of interest in the wellbore to collect formation fluid samples in storage chambers placed in the tool.
During drilling of a wellbore, a drilling fluid (“mud”) is used to facilitate the drilling process and to maintain a pressure in the wellbore greater than the fluid pressure in the formations surrounding the wellbore. This is particularly important when drilling into formations where the pressure is abnormally high. If the fluid pressure in the borehole drops below the formation pressure, there is a risk of blowout of the well. As a result of this pressure difference, the drilling fluid penetrates into or invades the formations for varying radial depths (referred to generally as invaded zones) depending upon the types of formation and drilling fluid used. The formation testing tools retrieve formation fluids from the desired formations or zones of interest, test the retrieved fluids to ensure that the retrieved fluid is substantially free of mud filtrates, and collect such fluids in one or more chambers associated with the tool. The collected fluids are brought to the surface and analyzed to determine properties of such fluids and to determine the condition of the zones or formations from where such fluids have been collected.
One feature that most formation testing tools have in common is a fluid sampling probe. This may consist of a durable rubber pad that is mechanically pressed against the rock formation adjacent the borehole, the pad being pressed hard enough to form a hydraulic seal. Through the pad is extended one end of a metal tube that also makes contact with the formation. This tube (“probe”) is connected to a sample chamber that, in turn, is connected to a pump that operates to lower the pressure at the attached probe. When the pressure in the probe is lowered below the pressure of the formation fluids, the formation fluids are drawn through the probe into the well bore to flush the invaded fluids prior to sampling. In some formation tests, a fluid identification sensor determines when the fluid from the probe consists substantially of formation fluids; then a system of valves, tubes, sample chambers, and pumps makes it possible to recover one or more fluid samples that can be retrieved and analyzed when the sampling device is recovered from the borehole.
It is desirable that only uncontaminated fluids are collected, in the same condition in which they exist in the formations. Commonly, the retrieved fluids are found to be contaminated by drilling fluids. This may happen as a result of a poor seal between the sampling pad and the borehole wall, allowing borehole fluid to seep into the probe. The mud cake formed by the drilling fluids may allow some mud filtrate to continue to invade and seep around the pad. Even when there is an effective seal, borehole fluid (or some components of the borehole fluid) may “invade” the formation, particularly if it is a porous formation, and be drawn into the sampling probe along with connate formation fluids.
U.S. Pat. No. 4,994,671 issued to Safinya et al. discloses a device in which visible and near infrared (IR) analysis of the fluids is performed in the borehole, without having to transport recovered samples of the fluid to the surface for chemical analysis. The infrared portion part of the electromagnetic spectrum (0.8 to 25 μm wavelength region, or equivalently wavenumbers of 12500 to 400 cm−1) of a substance contains absorption features due to the molecular vibrations of the constituent molecules. The absorptions arise from both fundamentals (single quantum level transitions occurring in the mid-infrared region from 2.5-25.0 microns) and combination bands and overtones (multiple quantum level transitions occurring in the mid- and the near-infrared region from 0.8-2.5 microns). The position (frequency or wavelength) of these absorptions contain information as to the types of molecular structures that are present in the material, and the intensity of the absorptions contains information about the amounts of the molecular types that are present. To use the information in the spectra for the purpose of identifying and quantifying either components or properties requires that a calibration be performed to establish the relationship between the absorbances and the component or property that is to be estimated. For complex mixtures, where considerable overlap between the absorptions of individual constituents occurs, such calibrations are accomplished using various chemometric data analysis methods.
In complex mixtures, each constituent generally gives rise to multiple absorption features corresponding to different vibrational motions. The intensities of these absorptions, to first order, will all vary together in a linear fashion as the concentration of the constituent varies. Such features are said to have intensities which are correlated in the frequency (or wavelength) domain. This correlation allows these absorptions to be mathematically distinguished from random spectral measurement noise which shows no such correlation. The linear algebra computations which separate the correlated absorbance signals from the spectral noise form the basis for techniques such as Principal Components Regression (PCR) and Partial Least Squares (PLS). As is well known, PCR is essentially the analytical mathematical procedure of Principal Components Analysis (PCA), followed by regression analysis. Second order effects include spectral changes resulting from changes in temperature or pressure. For example, these environmental factors change the intermolecular spacing and the degree to which the spectrum of one molecule is affected by the presence of its neighbors, such as the degree of hydrogen bonding. Nonlinear methods such as neural networks may also be used to correlate fluid composition and properties to the spectra.
PCR and PLS have been used to estimate elemental and chemical compositions and to a lesser extent physical or thermodynamic properties of solids, liquids and gases based on their mid- or near-infrared spectra. These chemometric methods involve: [1] the collection of mid- or near-infrared spectra of a set of representative samples; [2] mathematical treatment of the spectral data to extract the Principal Components or latent variables (e.g. the correlated absorbance signals described above); and [3] regression of these spectral variables against composition and/or property data to build a multivariate model. The analysis of new samples then involves the collection of their spectra, the decomposition of the spectra in terms of the spectral variables, and the application of the regression equation to calculate the composition/properties.
In Safinya et al. light the visible and near IR region is passed through the fluid sample. A spectrometer measures the spectrum of the transmitted and the back scattered light, and knowing the spectrum of the incident light, transmission and backscattered absorption spectra for the sample are determined. Using absorption spectra of water, gas, crude and refined oils, and drilling fluids, a least squares analysis is performed that models the observed spectra as a weighted sum of the spectra of its components, the least squares analysis giving the composition of the fluid in terms of weights of the various components.
Currently spectral analysis downhole and on site analysis for fixed single color interference filters is limited to around 11-30 nm full width at half maximum filters thus providing relatively low spectroscopic resolution. These filters are not suitable to distinguish between closely spaced spectral peaks or to identity isotopes whose spectral peak spacing's are much smaller than 11 nm. Thus, there is a need for an analysis technique suitable for downhole and onsite surface spectroscopic analysis of hydrocarbon samples with high resolution.